A Reliable Method for MRI Volumetric Assessment of Hippocampal Resection Following Anterior Temporal Lobectomy

Purpose: Surgical management of temporal lobe epilepsy in children is an established management of temporal lobe epilepsy (TLE). Surgery for TLE ranges from standard anterior temporal lobectomy (ATL) to selective resections which seek to spare temporal neocortex. Cortical dysplasia (CD) represents the most frequent non-neoplastic cause of TLE in children. Extent of resection (ER) of cortical dysplastic processes has been correlated with seizure control. The goal of this study was to design and test a standardized protocol based on anatomical boundaries for pre-operative (pre-op) and post-operative (post-op) imaging in order to measure hippocampal ER. Methods: Medical records and neuroimaging of a 2 year old developmentally delayed female TLE patient with CD in her right temporal lobe were reviewed. Seizure semiology, electrodiagnostics, and imaging studies indicated independent bilateral temporal dysfunction with the most severe deficit of activity in the right temporal lobe. Pathological analysis of temporal neocortex and mesial structures disclosed evidence of cortical dysplasia in all specimens. Pre-op MRI studies were performed using a Seimens Verio 3T unit generating 3 sequences: 1) T1 Magnetization Prepared Rapid Gradient Echo 3D with 1.0mm slice thickness, 2) T2 Turbo Spin Echo (TSE) Axial with 4.0mm slice thickness and, 3) T2 TSE Coronal with 3.0mm slice thickness. Post-op MRI studies were performed using a Seimens Espress 1.5T unit generating 3 sequences: 1) T1 Spin Echo (SE) Axial with 5.0mm slice thickness, 2) T2 TSE Axial with a 5.0mm slice thickness and, 3) T2 TSE Coronal with a 3.0mm slice thickness. Studies were co-registered via Hermes, a product of Hermes Medical Solutions, which rendered multiple image studies into one file. The resulting 3D brain map was then loaded into Amide, a free image viewing tool for registering and analyzing medical data image sets, which allowed a slice by slice evaluation of both pre-op and post-op images. After hippocampal boundaries were identified on each coronal slice, a region of interest (ROI) was created outlining hippocampal area. Amide calculated ROI areas for each slice and summed ROI areas to yield hippocampal volume. Results: Analysis of ROI yielded pre-op hippocampal volume of 2.691 cm3 vs. post-op volume of 0.259 cm3. ATL in our patient removed 96.25% of the hippocampus, leaving 3.75% of hippocampal tail remaining. Analysis completed after construction of protocol used to determine hippocampal anatomical borders. Conclusions: This project successfully designed and tested a standardized protocol based on anatomical boundaries for pre-op and post-op imaging in order to measure hippocampal volume ER. This protocol will allow for measurement of ER in future studies of TLE in pediatric populations.